seudomonas aeruginosa is a rod-shaped, Gram-negative, aerobic bacteria commonly isolated from clinical specimens.1 It is a major agent of healthcare-associated infections in pediatric intensive care units (PICUs), causing bloodstream infection, pneumonia, surgical site, and urinary tract infection.2 Its minimal nutritional requirements and its tolerance to difficult physical conditions contribute to its spread in the healthcare setting, making prevention and control difficult. Endemic reservoirs of P. aeruginosa in the hospital setting include sinks, flowers, and uncooked vegetables, reflecting its predilection for moist environments.1 Epidemics have been traced to respiratory equipment, cleaning solutions, sink drains, disinfectants, bronchoscopes, water systems, mouth swabs, bottled mineral water, and sometimes healthcare workers.1,3–6
In December 2006, the Infection Control Department of the University of Geneva Hospitals (HUG) was alerted to a possible increase in the number of P. aeruginosa isolates in patients hospitalized in the PICU. Between April and December 2006, 13 new cases of P. aeruginosa infection or colonization were attributed to the PICU, a number well above the normal range for the 2002–2005 period. The increase involved mainly urinary and tracheal isolates. Many of the affected patients were African children who had been transferred to Geneva to undergo cardiac surgery. We launched an investigation to describe and identify the cause of the outbreak and prevent additional cases.
The HUG is a large referral center that provides primary and tertiary medical care for Geneva, Switzerland and the surrounding area (approximately 800,000 population). Approximately 700 children are admitted annually to the 11-bed PICU (average duration of stay, 5.1 days; occupancy, 95%). The unit consists of 3 wards, each equipped with a sink.
In 2006, 150 children were admitted to the PICU after cardiac surgery. Approximately 60% of these patients are non-Swiss citizens supported by a humanitarian nongovernmental organization (NGO). Most cardiac surgery patients are admitted postoperatively to the same PICU ward (number 3) due to its greater capacity to accommodate space-consuming medical equipment such as hemofiltration (HF) and extracorporeal membrane oxygenation (ECMO) machines.
Many children undergoing cardiac surgery at HUG are transferred to Switzerland through a NGO dedicated to providing specialized treatment to children from low-income countries, mainly located in sub-Saharan Africa. Before and after surgery, most children are housed in a dedicated specialized residential center (NGO center) located 120 km from Geneva. This center encompasses 5 buildings and can host up to 54 children simultaneously. Children share meals and play together in a nonmedical environment. Approximately, 30 employees work in the center, including 4 nurses. This “home away from home,” which is independent from HUG, does not possess a dedicated infection control team, but the center's head nurse received training in infection control, and the center can get the assistance of infection control specialists from the neighboring hospital when needed.
Initial Outbreak Investigation/Epidemiologic Investigation
The outbreak investigation included 6 concurrent areas of inquiry: (1) assessment and reinforcement of infection control practices to prevent further transmission; (2) retrospective chart review; (3) environmental sampling; (4) prospective screening cultures of residents of the NGO center undergoing cardiac surgery; (5) screening cultures of residents and case-control study at the NGO housing facility to identify risk factors for acquisition; (6) molecular typing to assess clonality of strains.
Case Definition and Case Finding
To identify cases, to define baseline rates and to identify factors associated with disease, we reviewed microbiology records and patient charts. A “case” was defined as any patient newly colonized or infected with P. aeruginosa >24 hours after admission to the PICU (or within 2 weeks following discharge) between April 1, 2006 and September 30, 2008, based on culture of samples from any body site. Cases were identified by reviewing the computerized microbiology database and infection control surveillance records.
Observation of Suspected At-Risk Procedures
Patient care was reviewed by observing 2 cardiac surgical procedures and by interviewing operating theater personnel. Dialysis and extracorporeal circulation instruments setup and manipulation were also checked.
To detect potential P. aeruginosa carriers at admission, surveillance cultures were implemented. From January 2007 to June 2008, nasopharyngeal and tracheal aspirate samples were obtained from every resident of the NGO center undergoing cardiac surgery within 24 hours of admission to HUG. Specimens were obtained in the operating theater by the anesthesiologist following induction and before the incision. Every child living in the NGO center in October 2007 and March 2008 was screened for P. aeruginosa carriage through throat and rectal swabbing.
To identify a possible inanimate source of the outbreak,2 environmental specimens were obtained from moist areas in the operating theater, the PICU, and the NGO center. Specimens collected included sink, sink drains, faucets, refrigerators, and respiratory therapy equipment. Hemofiltration, ECMO machines, ventilators, stethoscopes, bronchoscopes, and gastroscopes were also inspected and cultured. Environmental cultures were performed using sterile cotton-tipped swabs (VWR Transport Swabs, Copan Italia, Brescia, Italy) premoistened with sterile saline. In addition, 10 to 20 mL specimen of liquids were collected in sterile jars from tap water, ice water, multidose medications, hand soaps, hand creams, alcohol-based handrub solutions, disinfectants, ultrasound gels, and cleaning solutions. Samples were immediately transported to the microbiology laboratory and plated within 4 hours of arrival. All liquids were filtered through 0.45-μm membrane filters and the filters were cultured on solid MacConkey agar. The nonfilterable fluids such as creams and lotions were directly inoculated into 100 mL nutrient broths.
To analyze the clonality of strains, all available clinical P. aeruginosa isolates from case patients, environmental, and surveillance cultures, and from the environmental sampling and child screening performed at the NGO center were typed and characterized by pulse-field gel electrophoresis (PFGE). DNA digestion and preparation was carried out with the use of SpeI (Bio-Rad, Hercules, CA) according to reagent kit instruction manual (Genepath, Bio-Rad). DNA separation was conducted by using the contour-clamped homogeneous electric field method on a contour-clamped homogeneous electric field Mapper (Bio-Rad) PFGE system. Isolates with identical restriction pattern were considered identical, those that differed by 3 or fewer bands were considered to be closely related, those that differed by 4 or 5 bands were considered possibly related, and those that differed by more than 6 bands were considered genetically different.4,7
We conducted an unmatched case-control study to assess the risk factors for acquisition of P. aeruginosa in children living in the NGO center. For the purpose of this study, cases consisted of every child living in the NGO center that was colonized or infected by P. aeruginosa through screening cultures performed in October 2007. Controls consisted of noncarriers concomitantly living in the facility. The following information was obtained for each child: age, sex, country of origin, underlying condition, length of stay, number of hospitalizations, presence of open wound or medical device, and exposure to antimicrobial therapy.
All data were entered in a spreadsheet and transferred to SPSS 15.0 (SPSS Inc, Chicago, IL) for analysis. We expressed continuous variables as the mean (±SD) or as the median and range if their distribution was skewed. For the case-control study, we performed Fisher exact tests to calculate P values. Multivariate analysis was not performed because of small sample size. In accordance with the ORION statement,8 with the exception of the univariate analysis of the case-control study, hypothesis testing for statistical significance was avoided.
The baseline rate of incident P. aeruginosa colonization or infection in the PICU for the 4 years preceding the outbreak was 7 cases per year (range, 5–10). There were no cases between January 2006 and the beginning of the outbreak in April 2006. During a 9-month period from April to December 2006, 13 patients with a median age of 3 years, 4 months (range, 10 days to 15 years) were identified as newly colonized or infected with P. aeruginosa more than 24 hours after admission to the PICU (Table 1). Most isolates (11/13) were uniformly susceptible to antipseudomonal agents. Four patients had bloodstream infection (2 primary bloodstream infections and 2 bloodstream infections secondary to pneumonia) and 1 had pneumonia without bloodstream infection. Eight patients were asymptomatic carriers (3 in the respiratory tract, 4 in the urinary tract, and 1 on an arterial catheter). The P. aeruginosa bloodstream infections were the first cases to occur in the PICU in more than 4 years; 3 of the 4 bacteremic patients died. Review of patients' charts revealed that 10 (76.9%) of the 13 attributable cases were young African recipients of a humanitarian NGO who had been transferred to Geneva to undergo surgery for an underlying congenital or acquired cardiopathy. In addition, 1 patient had undergone emergency thoracotomy during an orthopedic procedure, and 2 patients were premature neonates who did not undergo surgery. Nine (69%) patients had undergone cardiac surgery in the same operating theater, and 10 (77%) had been admitted to the same room in the PICU after surgery. Ten (77%) children required extracorporeal circulation at the time of surgery, and 3 (23%) patients had required HF while in the PICU.
The first working hypothesis was that patients had been contaminated at the time of surgery or during their stay in the PICU, perhaps due to contamination of the ECMO or HF equipment. After chart review, the very short time period (≤48 hours) between PICU admission and detection of colonization in 4 cases (all of them residents of the NGO center) was deemed atypical of healthcare-acquired P. aeruginosa, which usually requires a longer period to colonize patients.9 This finding led to the alternative hypothesis that patients were colonized before their admission and that the reservoir might be external to HUG. In consequence, prospective surveillance of cardiac surgery patients was initiated.
Between January 1, 2007 and September 30, 2008, preoperative screening identified 9 cases of P. aeruginosa colonization in NGO children on admission, representing 13% (9/67) of all NGO children admitted to the PICU during this period (Fig. 1). All 9 patients were asymptomatic carriers. Furthermore, on location screening of every child living at the NGO center was conducted twice. These investigations revealed that 38% (11/29) and 13% (7/54) of residents were throat and/or digestive carriers of P. aeruginosa in October 2007 and March 2008, respectively. One child, a 6-year-old boy with noma, also had a postenucleation infection of the orbital cavity caused by P. aeruginosa. During the same time period, 15 additional cases of P. aeruginosa infection or colonization were detected in PICU patients >24 hours after admission (Fig. 1). Five of these patients (30%) had lived at the NGO center before hospitalization but had negative preoperative screening, while 10 were children from the community. Most of these cases represented urinary tract colonization. Overall, NGO recipients represented 65% (24/37) of cases of P. aeruginosa infection or colonization during the whole outbreak period.
Seventy-three environmental specimens were collected at HUG. None of the specimens collected from the operating room, endoscopes, HF device, ECMO water, soaps, and skin care lotions yielded P. aeruginosa. One culture of the sink outlet located in ward number 3 in the PICU yielded P. aeruginosa. All other cultures from the PICU failed to show growth of P. aeruginosa.
In addition, a thorough inspection of the NGO center was conducted by the local health authorities in October 2007 in search of a potential reservoir. Fifty-two specimens were collected from the water system and from inanimate objects. None of the water, faucet, and showerhead samples and none of the soaps, skin lotions, and cleaning solutions grew P. aeruginosa. However, 10 of 14 sink and shower drains yielded P. aeruginosa, suggesting widespread colonization of the sewage system.
Because the microbiology laboratory routinely saves only organisms isolated from sterile body fluids, only 6 isolates (46%) were available for molecular typing before the detection of the outbreak in December 2006. Overall, 28 patient isolates were available for molecular analysis: 11 specimens from clinical samples, 6 from admission screening of cardiac surgery patients, and 11 from children living in the NGO center. In addition, the specimen from the PICU sink outlet and 10 environmental specimens from the NGO center were analyzed (Fig. Supplemental Digital Content 1, https://links.lww.com/INF/A233).
PFGE revealed 26 different patterns. A first group of closely related strains (group A) included isolates from 4 NGO patients who developed an infection while in the PICU, from a sink outlet in the PICU ward where these patients had been hospitalized, from 4 additional residents of the NGO center on their admission at HUG, and from a sink drain at the NGO center. An additional group of closely related strains (group L) included isolates from 5 children screened at the NGO center, including the young boy suffering from the postenucleation orbital infection. PFGE of the environmental samples retrieved from the NGO center revealed 10 distinct patterns.
In October 2007, 11 of 29 NGO center residents (38%) were colonized or infected with P. aeruginosa. They were compared with the 18 noncarriers (controls) in a case-control study to elucidate risk factors for P. aeruginosa acquisition at the NGO center (Table, Supplemental Digital Content 2, https://links.lww.com/INF/A234). The cases and controls were similar with respect to age, sex, and origin. The study did not show any significant difference in terms of underlying condition, length of stay, number of hospitalizations, presence of open wound or medical device, and exposure to antimicrobial therapy before the study day.
Termination of Outbreak and Long-Term Follow-Up
To control and terminate the outbreak, the following measures were undertaken. Standard precautions were reinforced; in particular, the importance of good hand hygiene practices was promoted. The PICU's contaminated sink and drain pipes were changed in February 2007 without prior chemical decontamination. Control cultures conducted 2 months after the replacement were negative. Two surgical procedures were observed and did not reveal significant breaches which could have explained the outbreak. Review of the ECMO and HF records and observation of the procedures did not reveal any irregularity. In addition, NGO recipients were screened for P.aeuruginosa before surgery, and empirical coverage for cardiac surgery patients presenting signs of sepsis was modified to ensure coverage of P.aeuruginosa. Preoperative antibiotic prophylaxis was not modified, however, and contact precautions of infected or colonized patients were not implemented.
Control measures undertaken at the NGO center included the reinforcement of hand hygiene practices and of proper nursing techniques (notably wound care). The use of common containers of lotions and creams was banned. In addition, removal of pipe deposits and disinfection of the sewage system using a hypochlorite solution were conducted in January 2008.
Between July 1, 2008 and December 31, 2008, a single case of P. aeruginosa infection occurred in the PICU. The patient, a 9-month old girl, had been transferred from Algeria to undergo emergency cardiac surgery. P. aeruginosa urinary tract colonization was detected on day 10. Molecular typing was not performed. However, the isolate was uniformly susceptible to antipseudomonal agents. As she was not a resident of the NGO center, admission screening was not performed.
We traced a polyclonal outbreak of P. aeruginosa infection and colonization in a PICU to a probable reservoir exogenous to the hospital. From a healthcare center perspective, this “outbreak” would be better qualified as a pseudoepidemic as the situation is secondary to an increase in the carriage rate of the admitted population. From a public health viewpoint, however, this outbreak represents a true epidemic. This report illustrates the importance of ruling out an external origin whenever an increase in the rate of detection of a pathogen is reported within a healthcare institution,3,10 and highlights the importance of close collaboration between infection control practitioners and public health authorities for the detection and containment of outbreaks. This report is the first to describe the occurrence of an ICU outbreak probably related to exogenous colonization of patients living in a residential center for specialized care.
Available evidence suggests that the probable reservoir was the water drainage system of the NGO center. It is hypothesized that children became colonized during their stay at the center and imported P. aeruginosa upon their admission to the PICU. The evidence supporting this hypothesis is substantial. First, environmental investigations at the NGO center showed widespread contamination of the drainage system, and a strain closely related to one of the outbreak strains was isolated from an environmental sample. Second, the very short delay between admission and colonization of some NGO children strongly suggests that patients were colonized before admission. Third, screening of NGO children admitted to HUG and children living at the NGO center showed a high (38%) prevalence of asymptomatic P. aeruginosa carriage. Some of these isolates were closely related to the PICU outbreak strain. In comparison, colonization rates of healthy persons in the community are usually much lower, less than 5% in most studies.1 The identification of the NGO center as the source of colonization of children is also supported by the fact that many shared closely related strains despite originating from different African countries. Thus, it appears unlikely that these children could have acquired similar strains in their home countries.
Extensive sampling of the hospital environment identified colonization of a single sink located in the PICU. ICU outbreaks related to contaminated water systems have been described.11 The colonized sink was probably not the point source, as its replacement failed to resolve the outbreak. Furthermore, local studies conducted using fluorescein did not reveal any significant back-splashing from this sink (data not shown).
It is unclear how children living in the NGO center acquired P. aeruginosa. The case-control study failed to identify any factors associated with P. aeruginosa acquisition while in the center. During their stay, children may have acquired the pathogen while washing with contaminated water or through back-splashing from a contaminated sink or shower drain, even though our investigation lacks definite evidence to support this hypothesis. Child-to-child transmission may also have occurred as they are encouraged to interact and play in close quarters. The presence of a child suffering from an active infection of the orbital cavity may have contributed to transmission. Child-to-child transmission has been documented in patients with cystic fibrosis during a winter camp.12 Cross-transmission through healthcare workers' hands may have also occurred. The decline in colonization rates between the October 2007 and March 2008 screenings surveys (38% vs. 13%) may suggest that the various decontamination procedures and control measures implemented at the NGO center were efficient to contain the outbreak.
In contrast to outbreaks that have been linked to the 2004 tsunami disaster,13 it should be stressed that the present outbreak was not caused by repatriates from foreign hospitals or the global migratory flow, but rather by a Swiss building which incidentally welcomes foreigners. Any individual living in this building, including Swiss citizens could have been contaminated with P. aeruginosa.
Several limitations of this investigation must be considered. Clearly, even though the origin of the outbreak can be traced to the NGO center, intrahospital cross-transmission probably played a role in the propagation of the outbreak in approximately one-third of cases. The acquisition of P. aeruginosa by 2 premature infants, for example, is undoubtedly attributable to the PICU. Also, we cannot exclude that the outbreak first started with the colonized sink in the PICU and that the colonized patients served as a vector to contaminate the NGO center's environment. Nevertheless, the outbreak was sustained by the NGO center, as shown by the carrier state of many patients before hospitalization in the PICU. Also, as patients underwent surgery in the same operating room and were hospitalized in the same ward, we cannot exclude that some cases were due to a reservoir within the hospital. Our extensive environmental investigation, however, does not support this hypothesis. Furthermore, only a fraction of positive clinical specimens were available for PFGE. Admission screening of non-NGO children was not conducted, so the baseline colonization rate of this population remains unknown. Finally, as the outbreak investigation revealed that not all strains were related, we cannot exclude that some of these children may have acquired P. aeruginosa while in contact with healthcare facilities in Africa.
The authors thank the personnel of the clinical microbiology laboratory (G. Renzi) for their laboratory assistance, the members of the department of anesthesiology for contributing to the prospective surveillance, and Rosemary Sudan for editorial assistance. The authors acknowledge the staff of the NGO center (C. Gutierrez), the PICU, and the departments of cardiac surgery and anesthesiology for their collaboration in the investigation of this outbreak.
1.Mandell GL, Bennett JE, Dolin R. Mandell, Douglas, and Bennett's Principles and Practice of Infectious Diseases
. 5th ed. Philadelphia, PA: Churchill Livingstone; 2000.
2.Long S. Principles and Practice of Pediatric Infectious Diseases
. 2nd ed. New York: Churchill Livingstone; 2003.
3.Mayhall CG. Hospital Epidemiology and Infection Control
. 3rd ed. Philadelphia: Lippincott Williams & Wilkins; 2004.
4.Foca M, Jakob K, Whittier S, et al. Endemic Pseudomonas aeruginosa
infection in a neonatal intensive care unit. N Engl J Med
5.Hota S, Hirji Z, Stockton K, et al. Outbreak of multidrug-resistant Pseudomonas aeruginosa
colonization and infection secondary to imperfect intensive care unit room design. Infect Control Hosp Epidemiol
6.Zawacki A, O'Rourke E, Potter-Bynoe G, et al. An outbreak of Pseudomonas aeruginosa
pneumonia and bloodstream infection associated with intermittent otitis externa in a healthcare worker. Infect Control Hosp Epidemiol
7.Tenover FC, Arbeit RD, Goering RV, et al. Interpreting chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis: criteria for bacterial strain typing. J Clin Microbiol
8.Stone SP, Cooper BS, Kibbler CC, et al. The ORION statement: guidelines for transparent reporting of outbreak reports and intervention studies of nosocomial infection. Lancet Infect Dis
9.American Thoracic Society; Infectious Diseases Society of America. Guidelines for the management of adults with hospital-acquired, ventilator-associated, and healthcare-associated pneumonia. Am J Respir Crit Care Med
10.Harbarth S, Pittet D. Multiresistance of gram-negative bacteria in intensive care units: bad news from without. Crit Care Med
11.Blanc DS, Nahimana I, Petignat C, et al. Faucets as a reservoir of endemic Pseudomonas aeruginosa
colonization/infections in intensive care units. Intensive Care Med
12.Ojeniyi B, Frederiksen B, Hoiby N. Pseudomonas aeruginosa
cross-infection among patients with cystic fibrosis during a winter camp. Pediatr Pulmonol
13.Uckay I, Sax H, Harbarth S, et al. Multi-resistant infections in repatriated patients after natural disasters: lessons learned from the 2004 tsunami for hospital infection control. J Hosp Infect